Thirteen experiments investigate the early development of the ability to learn patterns in sequential input. Such evidence would seem likely to provide a substantial contribution to our knowledge of how infants identify both concrete and abstract patterns, a fundamental aspect of cognitive development. The proposed experiments are organized into three broad studies. In the first study, we will examine "ancillary" factors that may influence sequence learning in infancy. We emphasize the importance of two kinds of relations between elements in any multi-element group or string: itemwise relations, between specific instances (e.g., statistical learning), and variablewise relations, between algebraic placeholders (e.g., abstract pattern or rule learning). We argue that a first step toward any programmatic series of experiments on these kinds of learning-must involve an understanding of processing limitations that may constrain performance. In the second study, we will investigate the different kinds of patterns infants may be able to learn, using past and pilot data as a guide for our theorizing. In the final study, we will explore the intriguing possibility that infants may learn more than one pattern in any single set of inputs, a classic question that has received little empirical attention in the literature on cognitive development. The short-term objectives of the present proposal are to elucidate fundamental developmental mechanisms in the ability to detect abstract visual and auditory structure. The long-term goals are to shed light on the larger question of how knowledge is acquired and structured in the human, how perceptual skills impact knowledge acquisition and structure, and how to best characterize early development. In the future, such understanding may aid in the formulation of diagnostics and treatments for some developmental disorders. [unreadable] [unreadable] [unreadable]